The rise of preprints in life sciences

Guest post by Christine Ferguson, 23/02/2018

If anything has gained traction in the open science arena over the past couple of years, it is preprints: the posting of research outputs as soon as they are in draft form, ahead of formal peer review. In this post, the first of two on the topic of preprints, I take a look back at noteworthy advances in the preprint arena.

A brief early history of preprints

ArXiv, the platform hosting preprints for physics, mathematics and computational biology, has been live since 1996 and currently hosts over 1.3 million preprints. This represents about 70% of the academic literature in these fields. The situation is not the same for the life sciences. There were early attempts in the 1960’s to push the idea in the domain of biological research, but the initiatives were quashed largely because journals refused to consider submissions that had previously been posted as preprints (Matthew Cobb provides an historical account). More recent initiatives for the life sciences failed to gain ArXiv-equivalent traction amongst researchers: Nature Precedings (2007-2012), PeerJ PrePrints (2013-current) and bioRxiv (2013-current).

In my previous life at PLOS Biology, discussions with authors before 2016 rarely featured preprints beyond offering reassurances that posting a preprint would not preclude a study from being published. However, enthusiasm began to warm about 2 years ago. Announcement followed announcement from organisations embracing preprints and advocating for their responsible and productive use. Interest from the research community, funders and the private sector for the support of preprints in life sciences brought ASAPbio to life (a California-based non-profit organisation focused on Accelerating Science and Publication in biology). Unlike ArXiv, there was no dominant server for the life sciences, so many publishers considered launching their own preprint servers. Alongside this, consultation by ASAPbio led to the proposal for a central aggregating service with regulation of disparate servers to be decided centrally.

2017, “The year of the preprint”

In January 2017, the UK-based MRC and Wellcome announced that they were going to accept preprints in grant applications and as products of grants. The NIH followed suit a few months later. A significant announcement came in April 2017 when the Chan Zuckerberg Initiative (CZI) announced a partnership with bioRxiv. The CZI pledged substantial funding to expand the bioRxiv server and its functionality with similar objectives to the proposed central service; ASAPbio subsequently relinquished plans to build that infrastructure. A snapshot of the preprints landscape as captured at that stage in 2017 shows the range of servers being launched, the date of launch and yet the modest numbers of preprints posted.

Courtesy of L. Page, Chief Innovation Officer at PLOS (May 2017)

These amount to no more than 1% of publicly-shared (published) outputs in the life sciences. At the close of 2017, dubbed ‘the year of the preprint‘ by Judy Luther, a chef in the Scholarly Kitchen, the future of preprints for biology was optimistic but far from settled.

The pros and cons of preprints

The concerns for the life sciences still outweigh the key advantages offered by preprints. The immediate sharing of results with the research community is, potentially, very beneficial to authors – although preprints need to pass a superficial screening and sanity check by server providers before posting, this is completed in days rather than the weeks required for the rigorous peer review process operated by journals. Furthermore, anyone in the community can contribute to and shape the study by submitting signed and public comments on the preprint server. The author(s) can, in turn, post improved versions of the preprint.

Not only can authors determine when their work is made public, but each posting is date-stamped, meaning authors have more control over establishing primacy for findings. That preprints are disseminated in the community more quickly than journal articles has obvious advantages for pandemic situations. However, researchers are wary of posting preprints for fear of being scooped and losing out on a potentially career-building publication in journals renowned for their impact factor. What we have seen as journal editors is that those authors who are willing to embrace posting a preprint do so only after submitting to a journal and receiving confirmation that formal peer review is underway.

Licensing how preprints may be distributed and reused is another issue for authors who are wary of permissive licenses. This is nicely summarised by Nature and touched on by this WIRED article. Subscription journals that use less permissive licenses cannot consider submissions that have been posted as preprints with CC-BY licenses. Perhaps a growth in CC-BY preprints will incentivise publishers to change, but, until then, this narrows the choice of journals for authors who want to ensure their preprint will ultimately be formally published.

Another issue is the cost of preprints: no cost is incurred by authors for posting or by readers for accessing preprints. However, the elephant in the room is that there is also no obvious business model for organisations that provide preprint servers. Therefore, although the preprint initiative promotes the ideals of open access and open peer review, the question of how to sustainably foot the bill for the infrastructure and submission screening process remains unresolved.

Open Research vs. Preprints

The advent of the ‘Open research’ publishing platforms is another arm of the preprint movement that caught on in 2017 and offers solutions to some of the concerns noted above. This is a publishing approach modelled on F1000Research, an open access journal offering immediate publication ahead of formal peer review. Unlike preprints, which are early drafts of research papers posted for public feedback but at the same time likely to be submitted to journals for formal peer review and publication, articles at F1000Research are formally published from the get-go. After this, a transparent version of formal peer review commences with reviewers selected by the authors. Ultimately, the initial publication is supplemented by the reviews (including the reviewers’ identities), author responses and any revised versions – each identifiable and citable via a DOI. Therefore, the author controls when to publish and who reviews their submission.

F1000 has partnered with and provided their publishing platform to a growing number of funders and forward-looking research institutions in the Open science arena: Wellcome was first to adopt the platform for research outputs generated by its grant-holders – Wellcome Open Research celebrated its first year of publishing in November 2017. Enthusiasm for ‘author-driven publishing’ caught on and the platform was quickly taken up in 2017 by the Bill & Melinda Gates Foundation, the Irish Health Research Board and the African Academy of Science for their grant-holders as well as by University College London (UCL) for affiliates engaged in child health research and the Montreal Neurological Institute for their research outputs. In return for managing the initial submission screening, posting, peer review and publication process, F1000 requires a small publishing fee which is based on word count, and will be covered in whole or in part by these funders and institutions. Notably, this fee is considerably lower than that charged by most other OA journals.

I’ll pause here to say that these early posting initiatives are promising experiments with clear traction amongst certain stakeholders – I am cautiously optimistic. In my next post, I will reflect on the impact that preprints are having on researchers and readers in 2018 and point to related opportunities arising for scholarly communication in the life sciences community.

About the author

Christine Ferguson (e-mail | Twitter | ORCID), Freelance life sciences editor and Open Science advocate and researcher

Trained lab biologist turned editor and Open Access journals professional, I was Joint Chief Editor at PLOS Biology (2014-2017). I am inspired by the principles of Open Science and the promise it holds for research and society, especially in resource-limited settings. Through consultancy work with stakeholders across the research and higher education community, my focus is to help identify and promote innovative technologies or practices that could lead to more effective knowledge production and exchange; and to understand the cultural change required for traction of these innovations.